Every scientific supercomputer powered by Nvidia GPUs will soon be a hybrid quantum computer, CEO Jensen Huang said during last week’s Nvidia GTC conference as he unveiled a quantum compute accelerator called NVQLink that connects “quantum and classical supercomputers — uniting them into a single, coherent system that marks the onset of the quantum-GPU computing era.” Accessible through Nvidia’s CUDA-Q software, NVQLink can create and test applications that draw on CPUs and GPUs alongside QPUs, “helping ready the industry for the hybrid quantum-classical supercomputers of the future.” At the conference, Huang also announced a Nokia partnership that will deliver AI-native 6G using new Nvidia tech.

Huang said the company developed NVQLink with input from the world’s leading supercomputing labs, and can let quantum processors talk to nine of them, including Brookhaven National Lab, Fermilab, the Lawrence Berkeley National Lab, Los Alamos National Laboratory and the MIT Lincoln Lab.

U.S. national laboratories, led by the Department of Energy, will use Nvidia NVQLink for quantum computing, the company explains in a newsroom post.

The Wall Street Journal writes that to be of practical use for businesses and researchers — at least in the near term — a hybrid approach is necessary, “in part because AI will be needed to perform full-scale error correction” for the notoriously unstable quantum computers, whose brains are Quantum Processing Units, or QPUs.

The announcement was made in a wide-ranging keynote speech by Huang that showcased a new platform called Aerial RAN Computer Pro (ARC-Pro) that aims to put his company at the heart of the 6G revolution as part of a $1 billion investment in Nokia.

Arc — an AI radio access network (AI RAN) computer — is being used by “telecom firms T-Mobile and Nokia to build ‘AI-native’ 6G cell phone towers, the next generation of wireless technology,” CNN writes.

Arc leverages an Nvidia hardware configuration — the Grace CPU, Blackwell GPU and purpose-build Mellanox ConnectX networking — to run Aerial software, a wireless communications system running atop CUDA-X that creates, for the first time, “a software-defined programmable computer” that will use AI in real time to improve spectral efficiency, Huang said.

Because Arc is compatible with Nokia’s existing 5G base stations it can “instantly upgrade millions of base stations around the world to 6G and AI,” according to Huang, who said  Nvidia has open-sourced the Aerial software to accelerate AI-native 6G.

Nvidia began talking-up AI-RAN in September. Data Centre Magazine reports that the Nokia partnership “marks a defining moment at the crossroads of datacenter architecture and telecommunications.”

Related:
Nvidia’s $400 Billion Week Fueled by Jensen Huang’s Dealmaking Spree, Bloomberg, 11/1/25
Nvidia Expands AI Ties with Hyundai, Samsung, SK, Naver, TechCrunch, 10/31/25